Regulation of gene expression in a cell is generally mediated by sequence-specific binding of gene regulatory molecules, often proteins, to chromosomal DNA. Regulatory proteins can effect either positive or negative regulation of gene expression. Generally, a regulatory protein will exhibit preference for binding to a particular binding sequence, or target site. Target sites for many regulatory proteins (and other molecules) are known or can be determined by one of skill in the art.
Despite advances in the selection and design of sequence-specific DNA binding gene regulatory proteins, their application to the regulation of an endogenous cellular gene can, in some cases, be limited if their access to the target site is restricted in the cell. Possible sources of restricted access could be related to one or more aspects of the chromatin structure of the gene. Access can be influenced by the structure of the gene per se (e.g., nucleotide methylation) or by the structure of the chromosomal domain in which the gene resides.
Cellular DNA, including the cellular genome, generally exists in the form of chromatin, a complex comprising nucleic acid and protein. Indeed, most cellular RNAs also exist in the form of nucleoprotein complexes. The nucleoprotein structure of chromatin has been the subject of extensive research, as is known to those of skill in the art. In general, chromosomal DNA is packaged into nucleosomes. A nucleosome comprises a core and a linker. The nucleosome core comprises an octamer of core histones (two each of H2A, H2B, H3 and H4) around which is wrapped approximately 150 base pairs of chromosomal DNA. In addition, a linker DNA segment of approximately 50 base pairs is associated with linker histone H1 (or a related linker histone in certain specialized cells). Nucleosomes are organized into a higher-order chromatin fiber (sometimes denoted a “solenoid” or a 30 nm fiber) and chromatin fibers are organized into chromosomes. See, for example, Wolffe “Chromatin: Structure and Function” 3rd Ed., Academic Press, San Diego, 1998 and Kornberg et al. (1999) Cell 98:285-294.
Chromatin structure is not static, but is subject to modification by processes collectively known as chromatin remodeling. Chromatin remodeling can serve, for example, to remove nucleosomes from a region of DNA, move nucleosomes from one region of DNA to another, change the spacing between nucleosomes or add nucleosomes to a region of DNA in the chromosome. Chromatin remodeling can also result in changes in higher order structure, thereby influencing the balance between transcriptionally active chromatin (open chromatin or euchromatin) and transcriptionally inactive chromatin (closed chromatin or heterochromatin).
Chromosomal proteins are subject to numerous types of chemical modification, some or all of which influence chromatin structure. For example, histones are subject to acetylation by histone acetyltransferases, deacetylation by histone deacetylases, methylation by histone methyltransferases (and therefore presumably to demethylation by histone demethylases), ubiquitination by ubiquitin ligases, de-ubiquitination by ubiquitin hydrolases, phosphorylation by histone kinases, dephosphorylation by histone phosphatases, and reversible ADP-ribosylation by poly-ADP ribose polymerase (PARP, also known as TFIIC). Strahl et al. (2000) Nature 403:41-45. Regulation of chromatin structure by methylation of histone H3 has been described. Rea et al. (2000) Nature 406:593-599. Modifications of non-histone chromosomal proteins include, for example, acetylation of HMG-1 (Munshi et al. (1998) Mol. Cell 2:457-467); HMGs 14 and 17 (Sterner et al. (1981) J. Biol. Chem. 256:8892-8895; Herrera et al. (1999) Mol. Cell. Biol. 19:3466-3473; Bergel et al. (2000) J. Biol. Chem. 275:11,514-11,520) and chromatin-resident transcriptional regulators such as, for example, TFIIE (Imhof et al. (1997) Curr. Biol. 7:689-692), p53 (Gu et al. (1997) Cell 90:595-606) and GATA-1 (Boyes et al. (1998) Nature 396:594-598). Chemical modification of histone and/or non-histone proteins is often a step in the chromatin remodeling process, and can have either positive or negative effects on gene expression. Generally, histone acetylation is correlated with gene activation; while deacetylation of histones is correlated with gene repression.
A number of enzymes capable of chemical modification of histones have been described and partially characterized. For example, histone acetyl transferases include Gcn5p, p300/CBP-associated factor (P/CAF), p300, CREB-binding protein (CBP), HAT1, TFIID-associated factor 250 (TAFII250), and steroid receptor coactivator-1 (SRC-1). Wade et al. (1997) Trends Biochem. Sci. 22:128-132; Kouzarides (1999) Curr. Opin. Genet. Devel. 9:40-48; Sterner et al. (2000) Microbiol. Mol. Biol. Rev. 64:435-459. The HDAC family of proteins have been identified as histone deacetylases and include homologues to the budding yeast histone deacetylase RPD3 (e.g., HDAC1, HDAC2, HDAC3 and HDAC8) and homologues to the budding yeast histone deacetylase HDA1 (e.g., HDAC4, HDAC5, HDAC6 and HDAC7). Ng et al. (2000) Trends Biochem. Sci. 25:121-126. The Rsk-2 (RKS90) kinase has been identified as a histone kinase. Sassone-Corsi et al. (1999) Science 285:886-891. A histone methyltransferase (CARM-1) has also been identified. Chen et al. (1999) Science 284:2174-2177.
Effects of alterations in chromatin structure upon gene expression have been reported or inferred. Fryer et al. (1998) Nature 393:88-91; and Kehle et al. (1998) Science 282:1897-1900.
Because of the dynamic structure of cellular chromatin, the ability of a regulatory molecule to bind its target site in a chromosome may be limited, in certain circumstances, by chromatin structure. For example, if a target site is present in “open” chromatin (generally thought of as nucleosome-free or having an altered nucleosomal conformation compared to bulk chromatin) structural barriers to the binding of a regulatory molecule to its target site are unlikely. By contrast, if a target site is present in “closed” chromatin (i.e. having extensive higher-order structure and/or close nucleosome spacing), steric barriers to binding are likely to exist. Thus, the ability of a regulatory molecule to bind to a target site in cellular chromatin will depend on the structure of the chromatin surrounding that particular target site. The chromatin structure of a particular gene can vary depending on, for example, cell type and/or developmental stage. For this reason, the regulation of a given gene in a particular cell can be influenced not only by the presence or absence of gene regulatory factors, but also by the chromatin structure of the gene.
Remodeling of chromatin can lead to activation of gene expression in vitro. For example, the NURF chromatin remodeling complex stimulates the transcriptional activation activity of the GAGA transcription factor. Tsukiyama et al. (1995) Cell 83:1011-1020. Transcriptional activation by a GAL4-VP16 fusion requires the RSF chromatin remodeling complex. LeRoy et al. (1998) Science 282:1900-1904. The SWI/SNF chromatin remodeling complex potentiates transcriptional activation by the VP16 activation domain and by ligand-bound glucocorticoid receptor. Neely et al. (1999) Mol. Cell 4:649-655; Wallberg et al. (2000) Mol. Cell. Biol. 20:2004-2013.
There are also several examples of a requirement for the activity of chromatin remodeling complexes for gene activation in vivo. The human SWI/SNF chromatin remodeling complex is required for the activity of the glucocorticoid receptor. Fryer et al. (1998) Nature 393:88-91. The mammalian SWI/SNF chromatin remodeling complex is required for activation of the hsp70 gene. de La Sema et al. (2000) Mol. Cell. Biol. 20:2839-2851. Mutations in the Drosophila ISWI protein adversely affect expression of the engrailed and Ultrabithorax genes. Deuring et al. (2000) Mol. Cell 5:355-365. Finally, mutations in the yeast SWI/SNF gene result in a decrease in expression of one group of genes and an increase in expression of another group of genes, showing that chromatin remodeling can have both positive and negative effects on gene expression. Holsteege et al. (1998) Cell 95:717-728; Sudarsanam et al. (2000) Proc. Natl. Acad. Sci. USA 97:3364-3369.
Despite this knowledge of the effects of chromatin remodeling on gene expression in vitro and in vivo, methods for directed manipulation of chromatin structure are not available. Accordingly, for situations in which a regulatory molecule is prevented, by chromatin structure, from interacting with its target site, methods for targeted modification of chromatin structure are needed. Such methods would be useful, for example, to facilitate binding of regulatory molecules to cellular chromatin and/or to facilitate access of DNA-binding molecules to cellular DNA sequences. This, in turn, would facilitate regulation of gene expression, either positively or negatively, by endogenous and exogenous molecules, and provide additional methods for binding these molecules to binding sites within regions of interest in cellular chromatin.